Re: Debug functionality of HTTPAPI: is there a way to suspend and then continue without starting anew?

James,

The source for HTTPAPI doesn't have any code to clear the debug file. You say there's no way to tell it not to clear it -- but, it NEVER clears it...   typically people will delete (unlink) it to make it start over...

It also has a process for you to provide your own debugging routines so you can do whatever you want with the debug data instead of writing it to the IFS -- which you could be using it make it work any way that you like.  (For example, putting the debug data into the job log, or continuing to write it to a stream file, but having your own logic to manage the file open/closes... it's all available to you.)

Extensive effort has been put into making it as versatile as possible.

FWIW -- most people do use separate files.

With regards to the password logging -- the password is sent over the network as part of the HTTP request chain -- so you can't turn it off via http_debug() without completely removing the HTTP request from the debug file -- which pretty much defeats the purpose of making a debug file.

As this is an RPG question, please consider asking on RPG400-L, or on the HTTPAPI forum at https://www.scottklement.com/forums/

-SK


On 7/21/23 12:38 PM, James H. H. Lampert via MIDRANGE-L wrote:

On 7/21/23 9:34 AM, Scott Klement wrote:

Why not use separate debug files for each program?

Under the present conditions, if there's no way to tell HTTPAPI not to clear the debug file, it appears that I have to check for problems (indicated via the U1 indicator), and save a copy of the debug file if they occurred, after *each* RPG program is called, instead of just once, at the end.

If I changed it to have each RPG program use a separate debug file, then (1) I'd have eight debug files instead of just one, and (2) I'd *still* have to check for problems, and save a copy of the debug file if they occurred, after each RPG program is called, instead of just once.

Also, turning it off for the purpose of hiding the password values doesn't make a lot of sense to me.

How would you, then, hide a password from the debug file, when it's passed as a parameter to the "signon" web service?

--
JHHL